A numerical method for portfolio selection based on Markov chain approximation

In this paper, A portfolio selection problem is approximated by a Markov chain which ismodulated by a continuous-time, finite-state, Markov chain. The main ingredient of theMarkov chain approximation is to approximate the wealth process and utility function oforiginal utility optimization problem by a controlled Markov chain. under the convergenceof the approximation scheme, Policy iteration methods as to obtain the optimal controls. Anumerical example is provided to illustrate the reability of the algorithm.

Keywords:

Numerical method , Portfolio selection , Stochastic optimal control , Markov chain approximation